Performance Prediction for Crop Irrigation Using Different Machine Learning Approaches

Performance Prediction for Crop Irrigation Using Different Machine Learning Approaches

Tarun Jain, Payal Garg, Pradeep Kumar Tiwari, Vamsi Krishna Kuncham, Mrinal Sharma, Vivek Kumar Verma
Copyright: © 2021 |Pages: 19
DOI: 10.4018/978-1-7998-7511-6.ch005
OnDemand:
(Individual Chapters)
Available
$37.50
Current Special Offers
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

Irrigation is an ancient practice that evolved over the years. Irrigation is the method through which a controlled amount of water is applied to plants making the most important recourse of irrigation. Earth is composed of 70% of water of which only 2.5% is fresh. Most of it trapped in snowfields and glaciers with only 0.007% of the earth's water for the needs of mankind. Limited water resources had become the main challenge in farming. In the chapter, machine learning algorithms and neural networks are used to reduce the usage of water in crop irrigation systems. This chapter focus on four mainstream machine learning calculations (KNN [k-nearest neighbor], GNB [Gauss Naive Bayes], SVM [support vector machine], DT [decision tree]) and a neural networks technique (artificial neural networks [ANN]) to expectation models utilizing the huge dataset (510 irrigation cases), bringing about productive and precise dynamic. The outcomes showed that k-nearest neighbors and artificial neural networks are the best indicators with the most elevated effectiveness of 98% and 90% respectively.
Chapter Preview
Top

I. Introduction

Agriculture played a vital role in human society endeavors attempting to be self-sufficient in food (Faye et al, 1998) in history. Irrigation is essential for crop production in many parts of the world. Various studies have shown the irrigation practices utilized in the country are lacking, mainly in the uniformity of water application is very low. The main reason is the lack of farmers skills to manage irrigation. This directly results in the wastage of water and the reduction of crop yields. The economy is mainly based on agriculture in a country like India and with isotropic climate conditions, yet farmers are unable to utilize agricultural resources. Various irrigation methodologies are used by our farmers such as manual irrigation using watering cans and buckets, drip irrigation, flood irrigation, sprinkler irrigation, etc. The present system has several limitations and one which is water wastage which can directly lead to water scarcity in drought areas and unhealthy production of crops.

Improving the farmer’s skill to manage and effectively control their irrigation system is as important as adopting the accurate irrigation scheduling methods. To address this problem, an automated irrigation system is developed by us where the irrigation takes place only when there is an acute requirement of water. Authors have made “PREDICTION OF CROP IRRIGATION SYSTEM”, a model for controlling and predicting irrigation facilities to help millions of farmers. Authors have made an effort to compare various machine learning algorithms and give detailed analysis on the performance of each on our dataset. Authors have shown that KNN works better for the dataset and can be used for improving the irrigation systems in India thereby helping the farmers to manage their crops easily without much skill.

Complete Chapter List

Search this Book:
Reset